Jobs

A job represents a single document extraction task. Upload a file, specify a schema, and ExtractForm processes it asynchronously.

Create a job (file upload)

curl -X POST http://localhost:4000/api/jobs \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -F "file=@document.pdf" \
  -F "type=EXTRACTION" \
  -F "schemaId=YOUR_SCHEMA_ID"

Job types

Type	Description
EXTRACTION	Extract structured data using a schema

Job statuses

Status	Description
PENDING	Created, awaiting queue
QUEUED	In processing queue
PROCESSING	Currently being processed
COMPLETED	Successfully finished
FAILED	Processing failed
CANCELLED	Cancelled by user

Presigned upload (large files)

For direct-to-storage uploads:

Step 1 — Get upload URL:

curl -X POST http://localhost:4000/api/jobs/upload-url \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"filename": "large-document.pdf"}'

Step 2 — Upload to the returned uploadUrl

Step 3 — Confirm and create job:

curl -X POST http://localhost:4000/api/jobs/confirm-upload \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "key": "RETURNED_KEY",
    "filename": "large-document.pdf",
    "type": "EXTRACTION",
    "schemaId": "YOUR_SCHEMA_ID"
  }'

Re-extract from existing document

curl -X POST http://localhost:4000/api/jobs/from-document/DOCUMENT_ID \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"type": "EXTRACTION", "schemaId": "YOUR_SCHEMA_ID"}'

Get job and extraction

# Job status and metadata
curl http://localhost:4000/api/jobs/JOB_ID \
  -H "Authorization: Bearer YOUR_TOKEN"

# Extraction result
curl http://localhost:4000/api/jobs/JOB_ID/extraction \
  -H "Authorization: Bearer YOUR_TOKEN"

Update extraction (manual edit)

Patch the extracted fields after review:

curl -X PATCH http://localhost:4000/api/jobs/JOB_ID/extraction \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"fields": {"invoice_number": "INV-001-CORRECTED"}}'

This emits a job.updated webhook event.

Cancel and restart

# Cancel a pending or queued job
curl -X POST http://localhost:4000/api/jobs/JOB_ID/cancel \
  -H "Authorization: Bearer YOUR_TOKEN"

# Restart a failed or completed job
curl -X POST http://localhost:4000/api/jobs/JOB_ID/restart \
  -H "Authorization: Bearer YOUR_TOKEN"

List jobs

curl "http://localhost:4000/api/jobs?page=1&limit=10&status=COMPLETED" \
  -H "Authorization: Bearer YOUR_TOKEN"

Query parameters: page, limit, type, status, parentRunId, schemaId.

Delete a job

curl -X DELETE http://localhost:4000/api/jobs/JOB_ID \
  -H "Authorization: Bearer YOUR_TOKEN"

Idempotency

Pass Idempotency-Key header on create endpoints to prevent duplicate jobs on retries. See Authentication.